A comprehensive skill for managing local LLM lifecycles, server reachability, and OpenAI-compatible API integrations within LM Studio.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install lm-studio
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install lm-studio using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is LM Studio?
The LM Studio skill empowers AI agents to take full control over local inference environments. It provides a structured framework for running local models, managing server availability, and ensuring that integrations remain privacy-focused and efficient. By using this extension for Openclaw Skills, developers can bridge the gap between local model execution and complex agentic workflows without relying on cloud-based providers.
This skill is designed for technical users who require fine-grained control over model loading, unloading, and resource allocation. It ensures that the AI agent can intelligently choose the right quantization and model size based on hardware limits, preventing common issues like out-of-memory errors while maintaining high-performance local APIs.
LM Studio Use Cases
- Running privacy-sensitive data processing through local models to ensure data never leaves the machine.
- Automating the lifecycle of local LLMs including downloading, loading, and swapping models based on current task needs.
- Developing and testing OpenAI-compatible applications using local endpoints before deploying to production.
- Debugging local inference performance by monitoring server reachability and resource usage.
- Integrating local models into MCP (Model Context Protocol) workflows for secure data retrieval and tool use.
How LM Studio Works
- The agent first verifies the local environment for LM Studio or llmster and ensures required utilities like curl and jq are present.
- It initializes a dedicated memory structure in the user's home directory to track model profiles and known-good server configurations.
- Before any integration task, the skill performs a reachability check to confirm the server is active on the expected port.
- It identifies the correct model identifier, distinguishing between downloaded files and active runtime instances.
- The agent executes a smoke test after every model or configuration change to validate the end-to-end request lifecycle.
- If local resources are exceeded or a task requires higher reasoning, the skill provides a clear path for escalation to cloud-based models.
LM Studio Setup
To enable local model management, install the skill using the following command:
clawhub install lm-studio
Ensure that LM Studio is installed and its server is configured to run on a known port (default is usually 1234). You can verify the initial setup by creating the memory directory:
mkdir -p ~/lm-studio/
LM Studio Data Schema & Taxonomy
The skill maintains state and configuration within a local directory to optimize subsequent sessions. The following table describes the organization of Openclaw Skills data for this module:
| Component | File Path | Description |
|---|---|---|
| Core Memory | ~/lm-studio/memory.md |
Tracks active ports, activation status, and preferred defaults. |
| Server Logs | ~/lm-studio/server-notes.md |
Documented reachability checks and server-mode configurations. |
| Model Index | ~/lm-studio/model-profiles.md |
A catalog of verified models categorized by specific task workloads. |
| Incident Log | ~/lm-studio/incidents.md |
A history of runtime failures and their confirmed technical resolutions. |
LM Studio Advanced Features
- Resource-Aware Model Matching: Automatically suggests smaller quantizations or shorter context lengths based on detected machine limits to prevent crashes.
- Multi-Step Validation: Executes specific smoke tests for completions, embeddings, and chat endpoints independently to ensure API stability.
- MCP Playbook Integration: Guided setup for Model Context Protocol servers to extend local model capabilities securely.
- Runtime State Synchronization: Distinguishes between listed, downloaded, and currently loaded models to prevent integration errors during runtime swaps.
- Local-First Guardrails: Enforces strict privacy by ensuring no data is routed to external networks unless explicitly requested by the user.
SKILL.md
Loading
Related Openclaw Skills
Auto: Task Automation & Scripting
A comprehensive automation engine for creating scripts, scheduling recurring jobs, and streamlining repetitive technical workflows.
Voice Communication
A professional integration for Telegram that provides automatic transcription of incoming audio and generates natural-sounding voice replies via TTS.
Agents Efficient Workflow
A high-efficiency coordination pattern that uses direct agent spawning and local file handoffs to minimize token consumption and context loss.
Review Analysis
A sophisticated diagnostic tool that converts unstructured customer reviews into prioritized action plans and pattern-based reports.
Comment Replies
An AI agent skill for drafting high-clarity, concise replies to social media comments and pre-sale objections.
Codex
Codex is a specialized skill for AI agents that enables safe, repository-aware coding with explicit sandbox controls and review-ready verification workflows.
Featured*
